Add the GPU rigid body pipeline from https://github.com/erwincoumans/experiments as a Bullet 3.x preview for Bullet 2.80

Extras/RigidBodyGpuPipeline/opencl/vector_add/VectorAddKernels.cl

@@ -0,0 +1,16 @@
+
+
+__kernel void VectorAdd(__global const float8* a, __global const float8* b, __global float8* c, int numElements)
+{
+    // get oct-float index into global data array
+    int iGID = get_global_id(0);
+	if (iGID>=numElements)
+		return;
+
+	float8 aGID = a[iGID];
+	float8 bGID = b[iGID];
+
+	float8 result = aGID + bGID;
+    // write back out to GMEM
+    c[iGID] = result;
+}